Abstract
Objective: The most recent diagnostic criteria for multiple sclerosis (MS) ascertain that findings from spinal cord MRI can be used to demonstrate dissemination in space. Because little is known about the prevalence and characteristics of cord lesions early in the disease, the authors studied the prevalence of spinal cord abnormalities in patients with early-stage MS and assessed their impact on diagnostic classification.
Methods: The brains and spinal cords of 104 recently diagnosed patients with MS were examined. Median interval between first symptom and diagnosis was 18.4 months. The brain MRI protocol included before and after gadolinium axial T1-weighted conventional spin-echo sequences and dual-echo spin-echo images. For spinal cord MRI, sagittal cardiac-triggered dual-echo T2-weighted and sagittal T1-weighted spin-echo images were included. Clinical assessment for each patient included age, sex, clinical signs for spinal cord involvement, and Expanded Disability Status Scale.
Results: Abnormal cord MRIs were found in 83% of patients, usually with only focal lesions. Diffuse cord abnormalities were found in 13% of patients, although in isolation they were found in only three patients. Focal cord lesions were often multiple (median number, 3.0), small (median, 0.8 vertebral segments), and primarily (56.4%) situated in the cervical spinal cord. In 68 of 104 patients (65.4%), two or more focal lesions were visible on spinal cord images. The criteria for dissemination in space, as defined in the McDonald criteria for the brain, were met in only 66.3% of the patients. This percentage increased to 84.6% when spinal cord MRI abnormalities were also included.
Conclusion: Spinal cord abnormalities are prevalent in patients with early-stage MS, have distinct morphologic characteristics, and help to determine dissemination in space at time of diagnosis.
MRI has become the major confirmatory test for multiple sclerosis (MS).1–3⇓⇓ However, a diagnosis, based on brain MRI alone, cannot be made with absolute certainty because a variety of diseases are known to cause MS-like white matter (WM) abnormalities in the brain. To increase MR specificity, diagnostic criteria were developed for MS patients based on the number, location, size, and enhancement of brain lesions.2–4⇓⇓ Although sensitive for MS, specificity of these MR criteria remains relatively low; even when applying criteria as proposed by Barkhof et al., specificity only reaches 73%.5
Given the high prevalence of spinal cord abnormalities in patients with established MS, it has been suggested that spinal cord MRI would be advantageous in the diagnostic workup of MS, both for establishing an early diagnosis and for recognizing clinical subtypes. Using phased array coils, the spinal cord can be imaged within a reasonable examination time and with high sensitivity for MS abnormalities.6–9⇓⇓⇓
Based on information obtained from spinal cord MRI, focal spinal cord lesions have been included in the latest MS MRI criteria as described by the McDonald panel.1 However, at the time of formulation of these McDonald criteria, no accurate data concerning the prevalence of spinal cord abnormalities at time of diagnosis were available.
Little is known about the prevalence of spinal cord abnormalities in the early stage of (clinically definite) MS. One study found spinal cord lesions in 42% of patients suspected of having MS.10 Although the positive predictive value of an abnormal cord MRI for MS was high, reported prevalence of MS spinal cord lesions was lower than in patients with established MS. Another study reported a higher prevalence of cord abnormalities (87.5%) in patients not yet converted to definite MS.11 In clinical practice, whether an MRI scan of the spinal cord is performed at the time of first presentation often depends on clinical presentation. For patients with an isolated cord syndrome, an MRI of the spinal cord is clearly indicated, whereas for patients with optic neuritis or brain stem syndrome, it is not generally recommended. To investigate the prevalence and characteristics of spinal cord MRI lesions in patients with early-stage MS, we focused on patients with recently diagnosed MS rather than on patients with “only” a clinically isolated syndrome because it excluded the possibility that the study would include patients who might not convert to MS.
We examined the prevalence and characteristics (number, size, location, and aspect) of spinal cord abnormalities in a large MS patient cohort, within only a few months of clinical diagnosis of MS. The spinal cord findings were then correlated and compared with brain MRI findings to define more accurately patients in whom spinal cord imaging might add relevant diagnostic information, according to the latest criteria for MS in MRI.1
Materials and methods.
Study group.
Brains and spinal cords of 104 patients with MS were examined within 6 months of clinical diagnosis (median time from diagnosis, 91 days). Diagnosis of MS was made based on evidence of dissemination in time and space, according to the criteria of Poser et al.,12 and was typically based on only clinical findings for 81 (78%) of the included patients and on a combination of clinical findings and laboratory results for the remaining 23 patients.
Patients were recruited from the MS Center of the University Medical Center and some surrounding nonacademic neurology outpatient clinics. All patients underwent assessment of disability using the Expanded Disability Status Scale (EDSS).13 Patients were classified into two groups according to disease course: relapse-onset MS and progressive-onset (primary progressive) MS. Informed consent was obtained from all participating patients.
MRI protocol.
MRI imaging was performed at 1.0 tesla (Magnetom impact, Siemens A.G., Erlangen, Germany) using a standard circularly polarized head coil and spinal phased-array coil. MRI protocol for the brain included before and after gadolinium axial T1-weighted conventional spin-echo (CSE) sequences (700/15/2 [repetition time (TR)/echo time (TE)/excitations]) and axial dual-echo CSE images (2,700/45 and 90/2). Slice position, number of slices (25), slice thickness (5 mm), interslice gap (0.5 mm), field of view (163 × 260 mm), and matrix (160 × 256) were identical for all MRI sequences. Gadolinium-DTPA (Gd-DTPA) was only administered when no contraindications or patient objections were present. Spinal cord MRI included a cardiac-triggered sagittal dual-echo CSE (2,400 to 2,900/20 and 80/1) and a sagittal T1-weighted CSE sequence (500/15/2, with gadolinium already onboard). Slice position, slice thickness (3 mm), interslice gap (0.3 mm), field of view (240 × 480), and matrix (256 × 512) were identical for both sequences.
Image analysis.
Examination results were printed on hard copy and scored by two experienced readers in consensus (F.B. and J.C.). Spinal cord abnormalities were scored regarding number, location, size, MR appearance (focal lesion or diffuse abnormality), and gadolinium enhancement. Furthermore, spinal cord images were scored for swelling or atrophy of the spinal cord at the site of a focal lesion. Focal lesions (i.e., sharply delineated areas of increased signal intensity [SI]) were considered to be present on CSE scans if seen on intermediate and T2-weighted MRIs. The size of focal lesions was expressed as their extension over a number of vertebral segments. Diffuse abnormalities were defined as areas with a subtle, poorly delineated increase of SI, best recognized as areas of SI higher than spinal CSF on intermediate-weighted images.14
T2-weighted brain MRIs were scored for number and location of focal lesions. T1-weighted MRIs were scored regarding number of hypointense lesions on precontrast images and number of enhancing focal lesions on postcontrast images. Focal WM abnormalities were defined as sharply demarcated areas of high SI on intermediate- and T2-weighted MRIs. Focal WM brain lesions were divided into four types based on their location: juxtacortical (contiguous with cortical gray matter), periventricular (adjacent to ventricles or at a distance of <1 cm from the ventricles, as measured from the center of the lesion), infratentorial (brainstem and cerebellum), and deep WM (remainder of WM lesions).
Statistical analysis.
Brain examinations were classified according to the diagnostic criteria defined by Paty2 and modified criteria by Barkhof et al.4,5⇓ The latter criteria require evidence of at least three of the following four criteria: 1) at least one gadolinium-enhancing lesion or nine T2 hyperintense lesions; 2) at least one infratentorial lesion; 3) at least one juxtacortical lesion; and 4) at least three periventricular lesions, and are used to determine dissemination in space in the context of the McDonald criteria.1 In addition, the diagnostic yield of combined brain and spinal cord MRI (where one missing brain lesion can be substituted for one spinal cord lesion) as proposed in the McDonald criteria was assessed.1 For normally distributed variables, differences between MS patient groups were assessed using a t-test for equality of means. For non-normally distributed variables, Mann–Whitney U tests were used, and for nominal variables, we used Fisher’s exact tests. Associations between MRI variables were assessed using Spearman rank correlation coefficient. For non-normally distributed data, median and interquartile range (IQR) are displayed. IQR is a measure of the spread of the data (the distance between the first quartile [25% percentile] and the third quartile [75% percentile] values).
Results.
Clinical findings.
One hundred four patients with MS (67 women and 37 men) were included (table 1). Of these 104 patients, 88 were classified as having relapse-onset MS, and 16 were classified as having progressive-onset MS. Mean age of the study group was 37.7 (± 9.1) years. From the time of the first presenting symptom until diagnosis, a median of 18.4 months (IQR = 45.7) elapsed. In general, patients presented low disability, and mean EDSS for the study group was 2.3 (± 0.97). Clinical signs of spinal cord involvement were found in 54 patients (52%), appearing most frequently in patients with progressive-onset MS (81%).
Table 1 Demographic and clinical descriptive by clinical subtype
Spinal cord.
Of all 104 spinal cord examinations, 86 (82.7%) were abnormal. In 73 of 104 patient cord examinations (70.2%), only focal lesions were observed. Diffusely involved spinal cords without focal lesions were seen in 3 of 104 patients (2.9%), and a combination of focal and diffuse abnormalities was seen in 10 of 104 patients (9.6%) (figure 1).
Figure 1. MRIs of a 42-year-old woman with multiple sclerosis with a relapse-onset disease course without clinical evidence of cord involvement and an Expanded Disability Status Scale score of 3. (Right) Sagittal intermediate-weighted (left) and T2-weighted (right) spinal cord images show a focal lesion at level C2 to C3 and diffuse abnormalities from C3 to C8. Adjacent cord images showed one additional focal lesion. (Left) Intermediate-weighted brain MRIs show (top) a periventricularly situated lesion, (middle) a juxtacortical lesion located in the temporal lobe, and (lower) an infratentorial lesion. More lesions were detected on brain images not shown. Brain images met Paty’s criteria, although Barkhof’s criteria were not met. McDonald criteria were met through substitution of a missing (periventricular) brain lesion with the focal cord lesion.
Patients with focally involved spinal cords mostly showed multiple small lesions (median, 3.0; IQR = 2.0) of a median of 0.8 segments (IQR = 0.3) (table 2 and figure 2). Focal lesions had an elongated configuration along the axis of the spinal cord. In 68 of 104 patients (65.4%), two or more focal lesions were visible on spinal cord images irrespective of the presence of diffuse abnormalities. Three hundred fifty-three focal cord lesions were detected, most of which were situated in the cervical spinal cord (56.4%), whereas lesions were least often detected at the level of the first six thoracic vertebrae (17.3%) (figure 3).
Table 2 Imaging findings by type of MS
Figure 2. MRIs of a 37-year-old man with multiple sclerosis with a relapse-onset disease course with clinical evidence of cord involvement and an Expanded Disability Status Scale score of 3. (Right) Sagittal intermediate-weighted (left) and T2-weighted (right) spinal cord images show focal lesions at the C1 to C2 and Th7 levels. Adjacent cord images showed three additional focal lesions situated in the cervical cord. (Left) Intermediate-weighted (top) and T2-weighted (bottom) brain MRIs show two periventricularly situated lesions (corpus callosum). Two infratentorial and two enhancing lesions were detected on brain images not shown. Brain images met Paty’s criteria, although Barkhof’s criteria were not met. McDonald criteria were met through substitution of a missing (periventricular) brain lesion with one focal cord lesion.
Figure 3. Distribution of focal lesions according to level in spinal cord of all patients with multiple sclerosis. Note the number of lesions is displayed as a percentage of the total number (n = 353).
Gadolinium was not administered in 27 patients who either refused IV injection or were known to have contrast allergy. In 15 of the remaining 87 patients (17.2%), gadolinium enhancement of focal cord lesions was observed. Median size of focal enhancing cord lesions was one segment (IQR = 0.5). In 11 of these 15 patients, focal enhancing lesions were located in cervical cord (see table 2).
Only 10 of the 353 focal lesions (2.8%) showed distention of the cord at the site of the lesion, 5 of which displayed contrast enhancement (figure 4). Focal spinal cord atrophy at the level of a focal lesion was observed in only 3 of the 353 focal lesions (0.8%) (figure 5), none of which showed enhancement. Only one patient showed decrease of SI on sagittal T1-weighted MRIs at the site where T2-weighted MRIs showed a focal lesion.
Figure 4. MRIs of a 33-year-old woman with multiple sclerosis (MS) with a relapse-onset disease course without clinical evidence of cord involvement and an Expanded Disability Status Scale score of 2. (Right) Sagittal intermediate-weighted (left), T2-weighted (middle), and gadolinium-enhanced T1-weighted spinal cord images show a focal, enhancing lesion at the C5 to C6 level. This lesion shows focal swelling of the cord at the site of the lesion, a rare characteristic for focal MS lesions. Also, note the small lesion in the medulla oblongata on intermediate- and T2-weighted images. (Left) Intermediate-weighted (top) and T2-weighted (bottom) brain MRIs show three juxtacortically situated lesions. More lesions, including two enhancing lesions, were detected on brain images not shown. Brain images met all MS criteria.
Figure 5. MRIs of a 41-year-old woman with multiple sclerosis (MS) with a progressive-onset disease course with clinical evidence of cord involvement and an Expanded Disability Status Scale score of 3. (Right) Sagittal intermediate-weighted (left) and T2-weighted (right) spinal cord images show two focal lesions. At the C2 level, a lesion shows focal atrophy of the cord, a characteristic that was only observed a few times. Also note the small lesion at the C7 to Th1 level. (Left) Intermediate-weighted (top) and T2-weighted (bottom) brain MRIs images show one infratentorial lesion situated in the right cerebellar hemisphere. More lesions, including two enhancing lesions, were detected on brain images not shown. Brain images showed 15 focal lesions, of which 13 were in typical locations, and all met MS MRI criteria.
Diffuse abnormalities were detected in 13 of 104 patients with MS; a mean of 11.2 (± 5.7) of a maximum 19 segments of spinal cord were diffusely involved (see table 2). Contour of the spinal cord, as seen on T2-weighted images at the site of detected diffuse abnormalities, did not show any local changes indicative of distension or atrophy (figure 6).
Figure 6. MRIs of a 41-year-old man with multiple sclerosis (MS) with a progressive-onset disease course with clinical evidence of cord involvement and an Expanded Disability Status Scale score of 4. (Right) Sagittal intermediate-weighted (left) and T2-weighted (right) spinal cord images. Note the diffuse spinal cord abnormalities at the cervical and high thoracic level on the intermediate-weighted image in contrast to the T2-weighted image, on which these abnormalities can hardly be depicted. Adjacent images showed continuation of the diffuse abnormalities throughout the entire spinal cord. (Left) Intermediate-weighted (top) and T2-weighted (bottom) brain MRIs show multiple periventricular lesions and one deep white matter lesion. More lesions, including one enhancing lesion, were detected on brain images not shown. Brain images showed 79 focal lesions, of which 19 were in typical locations, and all met MS MRI criteria.
Brain MRI findings.
In our series of 104 patients with MS, only two brain MRI examinations (1.9%) were normal. A median of 32 (IQR = 52.8) lesions was found (see table 2) with the following locations: 31.6% periventricular, 10.2% juxtacortical, 50.5% in deep WM, and 7.6% infratentorial. Of all focal lesions seen on T2-weighted MRIs, 26.5% appeared as hypointense lesions on T1-weighted MRIs (table 3). Enhancing lesions were seen in 39 of the 87 patients (44.8%) receiving gadolinium; these 39 patients had a median of three focal enhancing lesions (IQR = 5.0) (see table 2).
Table 3 Distribution (%) of brain lesions: no relation to spinal cord abnormalities
Relation between MS abnormalities in brain and spinal cord: whole group analysis.
The number of supratentorially located focal enhancing brain lesions correlated moderately with the presence of enhancing focal lesions in the spinal cord (r = 0.2; p = 0.02). For the remainder, there were no obvious associations between presence, type, and number of spinal cord abnormalities and number or location of brain lesions (see table 3), indicating that the occurrence of spinal cord abnormalities is largely independent of brain lesions.
Value of spinal cord MRI to brain MRI alone: brain criteria in combination with cord imaging.
Of all 104 examinations, 85 (81.7%) were abnormal (any abnormality) for brain and spinal cord examination. In 17 of the 104 patients (16.3%), results of brain examination were abnormal, whereas spinal cord examinations were normal. Only one patient’s brain examination was completely normal, whereas the spinal cord examination showed two focal lesions, and in another patient, brain and spinal cord examinations were completely normal. In 68 of the 104 patients (65%), two or more focal spinal cord lesions were present on MRI examination, and only 39 of these 68 (57%) patients showed clinical signs of spinal cord involvement.
Paty criteria for MS were met in 100 of the 104 brain examinations (sensitivity, 96.1%); for Barkhof/Tintore criteria (McDonald brain only), this was 69 of 104 (sensitivity, 67.3%) (table 4). In three of the four patients in whom brain examination was negative for Paty criteria, spinal cord MRI abnormalities were present; for patients negative for Barkhof/Tintore criteria, this was the case in 28 of the 35 examinations (see table 4). When a focal spinal cord lesion was substituted for one brain lesion to meet McDonald MRI criteria for dissemination in space, a sensitivity of 84.6% (88 of 104) was reached.1 If any number of focal cord lesions were substituted for any number of brain lesions, 98 of 104 patients would meet criteria of dissemination in space (sensitivity, 94%). Sensitivity of these last criteria decreased from 94% to 78% when using only lesions detected in the cervical (levels C1 to C7) spinal cord (p = 0.005).
Table 4 MR criteria for MS: additional value of spinal MR imaging to determine dissemination in space
Discussion.
In the diagnostic workup of a clinical setting, the majority of patients will only be referred for MRI examination (of the spinal cord) when symptoms indicate so. For many patients, this may not be when first symptoms appear or at the initial evaluation, certainly when the initial symptom is optic neuritis (possibly only seen by an ophthalmologist) or a sensory symptom (which may be recognized only in retrospect). Therefore, this study examined newly diagnosed patients with clinically definite MS and excluded patients with a clinically isolated syndrome. To assess the value of a spinal cord MRI in addition to brain MRI in this early diagnostic setting, we performed MRI examinations of the brain and spinal cord, irrespective of the clinical presentation. The mean age of the MS patient group was 37.7 years with low disability, as expressed by the EDSS (mean, 2.3), and might be described as typical for an early-stage MS patient group. There was a median interval of 18.4 months between first symptom and diagnosis.
The prevalence of cord abnormalities in our patient study population with newly diagnosed MS was high (83%). Studies describing patients with clinically isolated syndromes report a relatively low prevalence of cord abnormalities (42% and 27%), partly because of inclusion of patients who will never convert to MS.10,15⇓ In 77% of patients converting within 12 months to either clinically definite or probable MS, baseline cord images already showed cord lesions.15 This study and our current data are more in line with studies of patients with established MS, which have reported prevalence of cord abnormalities in up to 90% of patients.6–9,16–18⇓⇓⇓⇓⇓⇓ Differences between studies regarding patient selection criteria or MRI techniques used might account for part of the variation observed.
Spinal cords of MS patients in this study were affected by focal lesions in 70.2% of patients, diffuse abnormalities in 2.9% of patients, and a combination of focal lesions and diffuse abnormalities in 9.6% of patients. Spinal cords, when affected by focal MS lesions, showed mostly multiple lesions (median, 3), whereas the majority was present in the cervical cord. However, almost half of the lesions would have been missed if only this region would have been examined.
Besides high prevalence of cord abnormalities (83% of patients with newly diagnosed MS in this study had abnormal spinal cord examinations), the characteristic appearance of MS spinal cord abnormalities may also help to differentiate MS from other diseases19 when atypical brain lesions are found.11,20–23⇓⇓⇓⇓ Spinal cord abnormalities in patients with inflammatory disorders and cerebrovascular diseases have been reported to frequently differ from those seen in patients with MS. Findings such as leptomeningeal enhancement, large lesions extending over multiple segments and more than half the cord’s thickness, and hypointense lesions on T1-weighted MRIs are rarely reported in patients with MS. A hypointense lesion on T1-weighted spinal cord images was seen in only 1 of the 104 patients with MS, a feature more often described in those with Devic’s neuromyelitis optica. In this patient, brain examination showed multiple focal lesions on characteristic locations for MS, a combination not observed in those with Devic’s neuromyelitis optica.22,24–33⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓
Although our cohort of patients with newly diagnosed MS showed many brain lesions, in a substantial number of patients, brain MRI alone was, according to published criteria, insufficient for radiologic documentation of MS. The majority of these patients’ spinal cord examinations depicted additional MS-related abnormalities, which points to an important role for spinal cord MRI when brain MRI is negative or inconclusive for patients with clinically suspected MS. In line with previous studies, we found that MRI criteria for MS, as defined by Paty et al., were most sensitive (96%). To increase specificity and accuracy, the McDonald International Panel recommended criteria for brain MRI as proposed by Barkhof et al. and adapted by Tintore et al. From our cohort, 35 of 104 patients (34%) did not meet these criteria despite recently having been diagnosed with MS. Substituting one spinal cord lesion for one brain lesion, an option suggested in the McDonald criteria, decreased the number of patients who did not meet the proposed MRI criteria for dissemination in space to 16 of the 104 patients (corresponding to an increase of sensitivity to 85%). If the diagnostic criteria would have allowed for an unlimited substitution of brain lesions by spinal cord lesions, the sensitivity would have improved even further (to 94%, 98 of the 104 patients). This increase of sensitivity is caused by the depiction of additional MS abnormalities in the spinal cord. A strong decrease in sensitivity was found when substituting brain lesions with only cervically located spinal cord lesions, underlining the relevance of using phased-array coils for detection of MS lesions in the spinal cord. These spinal cord lesions appear to occur largely independently from brain lesions because no strong correlation was found between brain and cord abnormalities.
Because of our study design, we were unfortunately not able to calculate accuracy and false-positive values for the assessed MS criteria. As previously described, any increase of sensitivity of MS MRI criteria for brain is accompanied by a decrease in specificity and an increase of false-positive cases.5 However, because of inclusion of spinal cord lesions in the McDonald MRI criteria, this may not be the case because prevalence of spinal cord lesions in MS patients is high, even in a early diagnostic setting, and in other CNS diseases, it is found to be low.21,23⇓
The McDonald diagnostic criteria have been proposed to harmonize the diagnostic workup for patients with MS and to reach a consensus on how MRI should be used and interpreted specifically.1 Ideally, the specific MRI criteria recommended should facilitate an early diagnosis of MS in patients when clinical dissemination in space and time is not (yet) obvious and should provide a “perfect fit” with clinical data once clinical dissemination in space and time has occurred. A recent study showed that the first of these objectives has been met; recent studies demonstrated that for patients with clinically isolated syndromes suggestive of MS, repeated MRI of the brain increases the yield of MS diagnosis at 3 and 12 months when compared with clinical follow-up evaluation only.34,35⇓ Our study of patients with a recent clinical diagnosis of MS, the first to specifically address spinal cord imaging in the context of the new diagnostic criteria, shows that the fit between clinical and MRI classification can be significantly improved by performing spinal MRI in all patients who do not meet the MRI criteria of brain alone. We recommend a more liberal use of spinal cord lesions for diagnosis, endorsing unlimited substitution of brain lesions by spinal cord lesions.
Acknowledgments
The MS-MR Center is supported by the Stichting Vrienden MS Research. Joost C.J. Bot is supported by a grant from The Stichting Vrienden MS Research (no. 97-307 MS).
- Received February 17, 2003.
- Accepted September 11, 2003.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.